We use neuroendocrine strategies to study the neurobiology of major depression. Studies over three decades suggest that neuroendocrine dysregulation in depression is an index of the altered brain function during the illness. Limbic brain areas are viewed as producing "overdrive" of the hypothalamo-pituitary-adrenal (HPA) axis in depression. However, definite confirmation of this theory is lacking. The reason is that clinical HPA studies to date have relied on the indirect measures of plasma ACTH and cortisol to infer the extent of the central overdrive. Our data show that the pituitary gland in fact performs an impressive function of masking the central hyperactivity of the HPA axis. In our experiments the focus is on pituitary function in depression. We remove cortisol feedback using metyrapone (MET). The central dysregulation is then unmasked. In Experiment 1 we study the circadian rhythm of cortisol and ACTH. Pilot data indicate that under MET the central overdrive on the HPA axis is much more severe than was believed. Experiment 1 also may resolve persistent questions about adrenal cortical sensitivity to ACTH in depression. A further payoff from Experiment 1 will be to obtain the first unmasked information on the circadian organization of HPA activity in normal man and in depression. Our analytic strategy includes deconvolution methods of characterizing pulsatile hormone release over time. In Experiment 2, we study pituitary responsiveness to hCRF, using deconvolution methods to describe the early hormonal response phase. We document fast feedback effects of cortisol on ACTH responses, an effect that confounds current testing methods. We also unmask the pituitary with MET so as to eliminate both fast and delayed steroid feedback effects on the observed responsiveness of pituitary corticotroph cells to hCRF in depression.